Postgraduate medical course optimization method and system based on semantic reasoning

Abstract

The present application relates to the technical field of course optimization, and a postgraduate medical course optimization method and system based on semantic reasoning, comprising: obtaining a medical learning course set, constructing a learning course graph, optimizing the learning course graph to obtain an optimized learning course graph, sequentially extracting a target learner course set from multiple target learner course sets, formulating a personalized learning course set, obtaining an optimal recommended course sequence, obtaining an ideal learning course of a target learner, obtaining an execution frequency, if the execution frequency is greater than a preset execution frequency threshold, taking any one optimal recommended course in the optimal recommended course sequence as the ideal learning course, summarizing the ideal learning courses to obtain an ideal learning course set, and completing postgraduate medical course optimization based on semantic reasoning based on the ideal learning course set. The present application can improve the personalization and accuracy of course recommendation.

Description

Technical Field

[0001] This invention relates to the field of course optimization technology, and in particular to a method and system for optimizing postgraduate medical courses based on semantic reasoning. Background Technology

[0002] Semantic reasoning is a technique based on knowledge representation and logical reasoning used to derive new knowledge or conclusions from existing knowledge. It achieves the reasoning process by analyzing and understanding the semantics of language, thus enabling the processing of complex and semantically rich information. Optimization of postgraduate medical courses refers to the analysis, adjustment, and improvement of the curriculum system for medical postgraduates through scientific methods and technical means, in order to enhance the scientific rigor, rationality, and effectiveness of the courses, and to meet the personalized learning needs and career development goals of medical postgraduates.

[0003] Traditional curriculum design is typically based on fixed syllabi and teaching plans, making it difficult to meet the individualized learning needs of each graduate student. Furthermore, optimizing the curriculum for medical graduate students requires consideration of factors such as prerequisite relationships between courses, learners' prior knowledge, and the frequency of course updates. Therefore, improving the personalization and accuracy of course recommendations is a pressing technical challenge. Summary of the Invention

[0004] This invention provides a method for optimizing postgraduate medical courses based on semantic reasoning and a computer-readable storage medium. Its main purpose is to improve the personalization and accuracy of course recommendations, and promote the development of medical education and talent cultivation.

[0005] To achieve the above objectives, this invention provides a method for optimizing postgraduate medical courses based on semantic reasoning, comprising:

[0006] The medical graduate student database was identified, and a medical learning course set was obtained based on the medical graduate student database. The medical learning course set includes multiple medical learning courses.

[0007] A learning course graph is constructed based on the medical learning course set, and then optimized to obtain an optimized learning course graph.

[0008] Multiple target learner course sets were identified. Each target learner course set includes multiple target learner courses, and each target learner course set corresponds one-to-one with a learner.

[0009] Extract target learner course sets sequentially from multiple target learner course sets, and develop personalized learning course sets based on the target learner course sets;

[0010] Receive medical course optimization instructions, and obtain the optimal recommended course sequence based on the medical course optimization instructions, the optimized learning course map, and the personalized learning course set. The optimal recommended course sequence includes three optimal recommended courses.

[0011] Obtain the ideal learning courses for the target learners and determine whether the optimal recommended course sequence contains ideal learning courses.

[0012] If the optimal recommended course sequence does not contain an ideal learning course, then the ideal learning course is imported into the optimized learning course graph to obtain an updated learning course graph. The updated learning course graph is then used as the optimized learning course graph, and the process of obtaining the optimal recommended course sequence based on the medical course optimization instruction, the optimized learning course graph, and the personalized learning course set is returned.

[0013] The number of times the optimal recommended course sequence is obtained based on the medical course optimization instructions, the optimized learning course map, and the personalized learning course set is obtained, and the number of executions is compared with the preset execution number threshold.

[0014] If the number of executions exceeds the preset execution threshold, then any one of the best recommended courses in the optimal recommended course sequence will be taken as the ideal learning course.

[0015] The ideal learning courses are compiled to obtain the ideal learning course set, and the postgraduate medical courses are optimized based on semantic reasoning based on the ideal learning course set.

[0016] Optionally, constructing a learning course graph based on the medical learning course set includes:

[0017] Extract one medical learning course from the medical learning course set one by one, and perform the following operations on each extracted medical learning course:

[0018] Remove one of the extracted medical learning courses from the medical learning course set to obtain a simplified learning course set;

[0019] Extract one simplified learning course from the simplified learning course set in sequence. Based on the extracted medical learning courses and the extracted simplified learning courses, obtain the list of medical course learners and the list of simplified course learners. Based on the list of medical course learners and the list of simplified course learners, confirm the list of common learners and the list of all learners.

[0020] Course similarity is calculated based on the list of common learners and the list of all learners, where course similarity is the ratio of the number of common learners in the list of common learners to the number of all learners in the list of all learners.

[0021] The course rating sets and simplified course rating sets corresponding to medical learning courses are collected respectively. The average course rating and average simplified course rating are calculated based on the course rating sets and simplified course rating sets.

[0022] The modified cosine similarity is calculated based on the average course rating, the average simplified course rating, the course rating set, and the simplified course rating set.

[0023] The comprehensive similarity is obtained based on course similarity and modified cosine similarity. The comprehensive similarity is then summarized to obtain a comprehensive similarity set. It is then determined whether there is a comprehensive similarity in the comprehensive similarity set that is greater than the preset comprehensive similarity threshold.

[0024] If there is no comprehensive similarity greater than the preset comprehensive similarity threshold in the comprehensive similarity set, then return to the step of sequentially extracting medical learning courses from the medical learning course set until the medical learning course set is empty.

[0025] If there is a comprehensive similarity greater than the preset comprehensive similarity threshold in the comprehensive similarity set, then the medical learning courses and simplified learning courses corresponding to the comprehensive similarity will be integrated to obtain an associated learning course group.

[0026] By aggregating related learning course groups, a set of related learning course groups is obtained. A learning course graph is constructed based on the set of related learning course groups, which includes multiple learning courses.

[0027] Optionally, the formula for calculating the modified cosine similarity is as follows:

[0028]

[0029] Where X represents the modified cosine similarity, E a,i E represents the rating of the i-th learning course in the learning course rating set. b,j This represents the rating of the j-th simplified learning course in the simplified learning course rating set. This represents the average course rating. Let represent the average score of the simplified learning courses, n represent the number of scores of the learning courses in the learning course score set, m represent the number of scores of the simplified learning courses in the simplified learning course score set, i represent the index of the learning course score, and j represent the index of the simplified learning course score.

[0030] Optionally, optimizing the learning course graph to obtain an optimized learning course graph includes:

[0031] Perform the following operations on each course in the learning course map:

[0032] Retrieve course textbooks based on the course content, retrieve the research major name based on the course textbooks, retrieve the set of university names based on the research major name, and then retrieve a set of similar textbooks with the same research major name from the set of university names.

[0033] If the set of similar learning textbooks is empty, then the research major name is expanded to obtain an updated major name, and the updated major name is used as the research major name. Then, the step of retrieving the set of school names based on the research major name is returned until the set of similar learning textbooks is not empty.

[0034] If the set of similar learning textbooks is not empty, then a set of similar learning courses is obtained based on the set of similar learning textbooks, and each similar learning course in the set of similar learning courses is combined with the learning course to obtain an associated similar course group.

[0035] By summarizing and associating similar course groups, a set of associated similar course groups is obtained. Based on the set of associated similar course groups and the learning course graph, an optimized learning course graph is constructed.

[0036] Optionally, the step of developing a personalized learning curriculum based on the target learner's curriculum set includes:

[0037] Undergraduate majors are obtained from the target learner course set, and course grade groups are obtained from the target learner course set, wherein the course grade groups in the course grade group set correspond one-to-one with the learner courses in the target learner course set.

[0038] The course grade sets are classified into basic course grade sets, professional course grade sets, and practical course grade sets. The highest and lowest basic course grades are calculated based on the basic course grade sets.

[0039] The highest and lowest professional course grades are calculated based on the professional course grade set, and the highest and lowest practical course grades are calculated based on the practical course grade set.

[0040] The strengths and weaknesses of courses are determined based on the highest and lowest grades in basic courses, the highest and lowest grades in professional courses, the highest and lowest grades in practical courses, and the lowest grades in practical courses.

[0041] Professional data is obtained based on undergraduate majors and research directions. Cross-disciplinary judgment is performed on the professional data to obtain judgment data, which can be either cross-disciplinary or non-cross-disciplinary data.

[0042] If the data is determined to be cross-disciplinary data, then obtain the professional span value and determine whether the professional span value is greater than the preset professional span threshold. If the professional span value is greater than the preset professional span threshold, then develop a cross-disciplinary learning course set based on the advantageous and disadvantageous courses.

[0043] If the data is determined to be non-interdisciplinary, then a non-interdisciplinary learning course set is developed based on the strengths and weaknesses of the courses.

[0044] A set of interdisciplinary or non-interdisciplinary learning courses can be used as a personalized learning course set, which includes one or more personalized learning courses.

[0045] Optionally, obtaining the optimal recommended course sequence based on medical course optimization instructions, optimized learning course map, and personalized learning course set includes:

[0046] According to the medical course optimization instructions, courses to be recommended are extracted sequentially from the optimized learning course map. The course matching degree is calculated based on the personalized learning course set and the courses to be recommended. The course matching degree is then compared with the preset course matching degree threshold.

[0047] If the course matching degree is greater than the preset course matching degree threshold, then the course to be recommended will be used as the recommended course.

[0048] If the course matching degree is less than or equal to the preset course matching degree threshold, then return to the step of sequentially extracting recommended courses from the optimized learning course graph until all recommended courses in the optimized learning course graph have been extracted.

[0049] Summarize the recommended courses to obtain recommended course groups, and obtain the optimal recommended course sequence based on the recommended course groups.

[0050] Optionally, the step of calculating the course matching degree based on the personalized learning course set and the courses to be recommended includes:

[0051] Extract personalized learning courses from the personalized learning course collection, obtain personalized course knowledge point groups from the personalized learning courses, summarize the personalized course knowledge point groups, and obtain a personalized course knowledge point group set.

[0052] Obtain the knowledge point set of the courses to be recommended, calculate the overlap ratio of knowledge points based on the knowledge point set and the knowledge point group set of personalized courses, analyze the difficulty of the courses to be recommended, and obtain the difficulty value of the courses to be recommended.

[0053] The difficulty of each personalized learning course in the personalized learning course set is analyzed to obtain a set of personalized learning course difficulty values. The average personalized course difficulty value is then calculated based on the set of personalized learning course difficulty values.

[0054] The course difficulty difference is calculated based on the difficulty value of the course to be recommended and the average personalized course difficulty value. The course difficulty difference is the absolute difference between the difficulty value of the course to be recommended and the average personalized course difficulty value.

[0055] Using a pre-built NLP model, vector transformation is performed on both the recommended courses and the personalized learning course set to obtain the embedded vectors of the recommended courses and the set of embedded vectors of the personalized courses. In the set of embedded vectors of the personalized courses, there is a one-to-one correspondence between the embedded vectors of the personalized courses and the personalized learning courses.

[0056] The average semantic distance is calculated based on the embedding vectors of the courses to be recommended and the set of embedding vectors of personalized courses. The course synergy effect value between the courses to be recommended and the set of personalized learning courses is calculated based on the average semantic distance, the courses to be recommended and the set of personalized learning courses, and the course moderating factor is obtained.

[0057] The course matching degree is calculated based on the overlap ratio of knowledge points, the difference in course difficulty, the course synergy effect value, and the course adjustment factor.

[0058] Optionally, obtaining the course adjustment factor includes:

[0059] Obtain the knowledge point sets of the advantageous courses and the disadvantageous courses. Calculate the similarity of the first course based on the knowledge point sets of the advantageous courses and the similarity of the second course based on the knowledge point sets of the disadvantageous courses.

[0060] The first course similarity is obtained by mapping the similarity using a pre-constructed logical function and a preset molecular adjustment factor;

[0061] The second course similarity is mapped using the aforementioned logical function to obtain the second mapping similarity. The course adjustment factor is calculated based on the first mapping similarity and the second mapping similarity, wherein the course adjustment factor is the product of the first mapping similarity and the second mapping similarity.

[0062] Optionally, the formula for calculating the course matching degree is as follows:

[0063]

[0064] Where M represents the course matching degree, The Cth value represents the set of difficulty values ​​for personalized learning courses. r Each personalized learning course difficulty value, H represents the number of personalized learning course difficulty values ​​in the set, and d c Z represents the difficulty level of the course to be recommended, T represents the overlap ratio of knowledge points, Q represents the synergistic effect of the course, and Q represents the course adjustment factor. C represents the difference in course difficulty. r This represents an index representing the difficulty values ​​of personalized learning courses.

[0065] To achieve the above objectives, the present invention also provides a postgraduate medical course optimization system based on semantic reasoning, comprising:

[0066] The course graph construction module is used to identify the medical postgraduate database, obtain the medical learning course set based on the medical postgraduate database, where the medical learning course set includes multiple medical learning courses, construct a learning course graph based on the medical learning course set, and optimize the learning course graph to obtain an optimized learning course graph.

[0067] The personalized learning course development module is used to identify multiple target learner course sets. Each target learner course set includes multiple target learner courses, and each target learner course set corresponds one-to-one with a learner. The target learner course sets are extracted sequentially from the multiple target learner course sets, and personalized learning course sets are developed based on the target learner course sets.

[0068] The optimal recommended course acquisition module is used to receive medical course optimization instructions, obtain the optimal recommended course sequence based on the medical course optimization instructions, the optimized learning course map, and the personalized learning course set. The optimal recommended course sequence includes three optimal recommended courses. The module also obtains the ideal learning courses for the target learner and determines whether there are ideal learning courses in the optimal recommended course sequence.

[0069] The course optimization completion module is used to import ideal learning courses into the optimized learning course graph if the optimal recommended course sequence does not contain ideal learning courses, thereby obtaining an updated learning course graph. This updated learning course graph is then used as the optimized learning course graph. The module returns to the step of obtaining the optimal recommended course sequence based on the medical course optimization instruction, the optimized learning course graph, and the personalized learning course set. It also retrieves the execution count of obtaining the optimal recommended course sequence based on the medical course optimization instruction, the optimized learning course graph, and the personalized learning course set, comparing the execution count with a preset execution count threshold. If the execution count is greater than the preset threshold, any one of the optimal recommended courses in the optimal recommended course sequence is taken as the ideal learning course. The ideal learning courses are then aggregated to obtain an ideal learning course set. Finally, the module completes the semantic reasoning-based optimization of postgraduate medical courses based on this ideal learning course set.

[0070] To address the above problems, the present invention also provides an electronic device, the electronic device comprising:

[0071] Memory, storing at least one instruction;

[0072] The processor executes the instructions stored in the memory to implement the semantic reasoning-based postgraduate medical course optimization method described above.

[0073] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one instruction, which is executed by a processor in an electronic device to implement the aforementioned method for optimizing postgraduate medical courses based on semantic reasoning.

[0074] To address the problems described in the background section, this invention identifies a medical graduate student database and obtains a set of medical learning courses based on this database. This set of medical learning courses includes multiple medical courses. By identifying the medical graduate student database, this invention integrates scattered student learning information and course information into a unified data source. A learning course graph is constructed based on the medical learning course set and then optimized to obtain an optimized learning course graph. This invention's learning course graph graphically presents the semantic relationships between medical learning courses, making the complex relationships between courses intuitive and easy to understand. Through optimization of the learning course graph, potential semantic connections between courses can be uncovered. This invention discovers previously unnoticed course combinations or knowledge expansion paths, providing new insights for course optimization and adjustment. It identifies multiple target learner course sets, each containing multiple target learner courses, with a one-to-one correspondence between the target learner course set and the learner. This ensures that each target learner course set accurately reflects the courses each learner has already studied or plans to study, providing a foundation for developing personalized learning course sets for each learner. The invention sequentially extracts target learner course sets from multiple target learner course sets and develops personalized learning course sets based on these sets. This invention develops personalized learning course sets based on each learner's target learner course set, enabling... By fully considering learners' existing knowledge and learning goals, and tailoring suitable learning courses for them, this invention helps improve learners' learning interest and effectiveness, making the courses more aligned with their actual needs. Receiving medical course optimization instructions, the invention obtains an optimal recommended course sequence based on these instructions, an optimized learning course map, and a personalized learning course set. This optimal recommended course sequence includes three optimal recommended courses. By combining the optimized learning course map and the personalized learning course set, this invention obtains an optimal recommended course sequence, providing learners with course recommendations that best match their needs and the logic of the curriculum system. This improves the accuracy and effectiveness of course recommendations, identifies the ideal learning courses for the target learner, and determines whether an ideal optimal recommended course sequence exists. This invention, by determining whether an ideal learning course exists for the target learner within the optimal recommended course sequence, ensures that the recommended courses meet the learner's personalized expectations. If no ideal learning course exists within the optimal recommended course sequence, the ideal learning course is imported into an optimized learning course graph to obtain an updated learning course graph. This updated learning course graph is then used as the optimized learning course graph, returning to the step of obtaining the optimal recommended course sequence based on the medical course optimization instructions, the optimized learning course graph, and the personalized learning course set. This invention, through continuous updating and optimization of the course graph, improves the accuracy and adaptability of course recommendations, ensuring that the recommended courses better meet the learner's expectations, continuously optimizing the course recommendation results, and increasing learner satisfaction.The invention obtains the execution count of the optimal recommended course sequence obtained based on medical course optimization instructions, optimized learning course graph, and personalized learning course set. The execution count is compared with a preset execution count threshold. This threshold prevents the algorithm from entering an infinite loop when an ideal learning course cannot be found, improving efficiency and stability. If the execution count exceeds the preset threshold, any optimal recommended course in the optimal recommended course sequence is selected as the ideal learning course. These ideal learning courses are then aggregated to obtain an ideal learning course set. Based on this set, semantic reasoning-based optimization of postgraduate medical courses is performed. This invention improves the quality and applicability of postgraduate medical courses, making them more aligned with students' learning needs and professional development requirements, thus contributing to the cultivation of more outstanding medical professionals. Therefore, this invention can improve the personalization and accuracy of course recommendations, promoting the development of medical education and talent cultivation. Attached Figure Description

[0075] Figure 1 A flowchart illustrating a method for optimizing postgraduate medical courses based on semantic reasoning, provided in an embodiment of the present invention.

[0076] Figure 2 A functional block diagram of a postgraduate medical course optimization system based on semantic reasoning provided in an embodiment of the present invention;

[0077] Figure 3 A schematic diagram of the structure of an electronic device for implementing the semantic reasoning-based postgraduate medical course optimization method according to an embodiment of the present invention;

[0078] Figure 4 This is a learning course map provided as an embodiment of the present invention for implementing the semantic reasoning-based graduate medical course optimization method.

[0079] Explanation of reference numerals in the attached figures:

[0080] 10. Electronic device; 11. Processor; 12. Memory; 13. Bus.

[0081] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0082] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0083] This application provides a method for optimizing postgraduate medical courses based on semantic reasoning. The executing entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0084] Reference Figure 1 The diagram shown is a flowchart illustrating a method for optimizing postgraduate medical courses based on semantic reasoning, according to an embodiment of the present invention. In this embodiment, the method for optimizing postgraduate medical courses based on semantic reasoning includes:

[0085] S1. Identify the medical graduate student database and obtain the medical learning course set based on the medical graduate student database. The medical learning course set includes multiple medical learning courses.

[0086] It should be explained that the medical graduate student database refers to a database containing information related to medical graduate students, which stores data such as their learning records, course selections, grades, and research directions.

[0087] S2. Construct a learning course map based on the medical learning course set, and optimize the learning course map to obtain an optimized learning course map.

[0088] In detail, the construction of the learning curriculum map based on the medical learning curriculum set includes:

[0089] Extract one medical learning course from the medical learning course set one by one, and perform the following operations on each extracted medical learning course:

[0090] Remove one of the extracted medical learning courses from the medical learning course set to obtain a simplified learning course set;

[0091] Extract one simplified learning course from the simplified learning course set in sequence. Based on the extracted medical learning courses and the extracted simplified learning courses, obtain the list of medical course learners and the list of simplified course learners. Based on the list of medical course learners and the list of simplified course learners, confirm the list of common learners and the list of all learners.

[0092] Course similarity is calculated based on the list of common learners and the list of all learners, where course similarity is the ratio of the number of common learners in the list of common learners to the number of all learners in the list of all learners.

[0093] The course rating sets and simplified course rating sets corresponding to medical learning courses are collected respectively. The average course rating and average simplified course rating are calculated based on the course rating sets and simplified course rating sets.

[0094] The modified cosine similarity is calculated based on the average course rating, the average simplified course rating, the course rating set, and the simplified course rating set.

[0095] The comprehensive similarity is obtained based on course similarity and modified cosine similarity. The comprehensive similarity is then summarized to obtain a comprehensive similarity set. It is then determined whether there is a comprehensive similarity in the comprehensive similarity set that is greater than the preset comprehensive similarity threshold.

[0096] If there is no comprehensive similarity greater than the preset comprehensive similarity threshold in the comprehensive similarity set, then return to the step of sequentially extracting medical learning courses from the medical learning course set until the medical learning course set is empty.

[0097] If there is a comprehensive similarity greater than the preset comprehensive similarity threshold in the comprehensive similarity set, then the medical learning courses and simplified learning courses corresponding to the comprehensive similarity will be integrated to obtain an associated learning course group.

[0098] By aggregating related learning course groups, a set of related learning course groups is obtained. A learning course graph is constructed based on the set of related learning course groups, which includes multiple learning courses.

[0099] It should be explained that a medical learning course refers to a single medical course within the medical learning course set. A simplified learning course refers to a course extracted from the simplified learning course set. The simplified learning course set is the subset obtained by progressively removing extracted medical learning courses from the medical learning course set. This invention extracts only one medical learning course from the medical learning course set at a time. This method allows the system to perform detailed analysis and processing on the currently extracted course at each step, rather than processing all courses at once. This ensures that each course receives sufficient attention, avoiding inaccuracies or omissions due to excessive data volume. Furthermore, extracting only one medical learning course at a time allows for a more accurate calculation of the similarity between that course and each simplified learning course in the simplified learning course set. Through progressive extraction and calculation, the system can analyze the relationships between courses in greater detail, thereby constructing a more accurate learning course graph. For example, a medical learning course set is (Medical Learning Course 1, Medical Learning Course 2, Medical Learning Course 3). If Medical Learning Course 1 is extracted from the set, then after removing Medical Learning Course 1 from the set, the resulting medical learning course set is (Medical Learning Course 2, Medical Learning Course 3), and this set is used as the simplified learning course set. The medical course learner list refers to the list of all students studying a particular medical learning course. The simplified course learner list refers to the list of all students studying the simplified learning course. The co-learner list refers to the set of all students simultaneously studying two courses. The all learner list refers to the set of all students studying two courses. The learning course rating set refers to the list of ratings from all students studying a particular course. The simplified learning course rating set refers to the list of ratings from all students studying the simplified learning course. The average learning course rating refers to the average of all ratings for a particular course. The average simplified learning course rating refers to the average of all ratings for a particular simplified learning course. The scoring described in this embodiment of the invention refers to the scores in both the learning course scoring set and the simplified learning course scoring set, which are the evaluation scores of each learner for that course in the past. Optionally, student evaluation scores for each course can be collected through questionnaires. The comprehensive similarity is the value obtained by adding the course similarity and the modified cosine similarity. The comprehensive similarity set is the set of all comprehensive similarities. The comprehensive similarity threshold is a pre-set value used to determine whether two courses are sufficiently similar, thereby deciding whether to link them. The associated learning course group refers to the medical learning courses and simplified learning courses whose comprehensive similarity is greater than the pre-set comprehensive similarity threshold. The associated learning course group set is the set of all associated learning course groups.The construction of a learning course graph based on a set of associated learning course groups refers to treating each learning course in the set as a node, obtaining all nodes, connecting each associated learning course group in the set to form edges, obtaining all edges, and constructing the learning course graph based on all nodes and all edges. Integration refers to the operation of placing medical learning courses and simplified learning courses into a set.

[0100] In detail, the formula for calculating the modified cosine similarity is as follows:

[0101]

[0102] Where X represents the modified cosine similarity, E a,i E represents the rating of the i-th learning course in the learning course rating set. b,j This represents the rating of the j-th simplified learning course in the simplified learning course rating set. This represents the average course rating. Let represent the average score of the simplified learning courses, n represent the number of scores of the learning courses in the learning course score set, m represent the number of scores of the simplified learning courses in the simplified learning course score set, i represent the index of the learning course score, and j represent the index of the simplified learning course score.

[0103] It should be explained that the modified cosine similarity is measured by calculating the cosine similarity of the rating vectors of the two courses and then centering the rating vectors (subtracting the average rating).

[0104] In detail, the optimization of the learning course graph to obtain an optimized learning course graph includes:

[0105] Perform the following operations on each course in the learning course map:

[0106] Retrieve course textbooks based on the course content, retrieve the research major name based on the course textbooks, retrieve the set of university names based on the research major name, and then retrieve a set of similar textbooks with the same research major name from the set of university names.

[0107] If the set of similar learning textbooks is empty, then the research major name is expanded to obtain an updated major name, and the updated major name is used as the research major name. Then, the step of retrieving the set of school names based on the research major name is returned until the set of similar learning textbooks is not empty.

[0108] If the set of similar learning textbooks is not empty, then a set of similar learning courses is obtained based on the set of similar learning textbooks, and each similar learning course in the set of similar learning courses is combined with the learning course to obtain an associated similar course group.

[0109] By summarizing and associating similar course groups, a set of associated similar course groups is obtained. Based on the set of associated similar course groups and the learning course graph, an optimized learning course graph is constructed.

[0110] It should be explained that "course textbooks" refers to the textbooks required for each course. "Research major name" refers to the name of the graduate-level professional field to which the course textbooks belong. "Institution name set" refers to the collection of all institutions offering courses in this research major. "Similar course textbook set" refers to the textbooks used for courses with the same research major name offered at different institutions. "Expanding the research major name" means expanding the research major name to a broader range of major names. "Updating the major name" means using the expanded major name as the new research major name. For example, if the research major name is "Ophthalmology (major code 100212) under the branch of Clinical Medicine (major code 1002)," expanding the research major name results in the updated major name "Clinical Medicine (major code 1002)." "Associated similar course group" is the combination obtained by combining each similar course in the similar course set with the course mentioned above, when the similar course set is not empty. "Associated similar course group set" refers to the collection of all associated similar course groups. The construction of an optimized learning course graph based on associated similar course groups and learning course graphs refers to the following: If some courses have already established associations in the learning course graph, and further associations are established in the associated similar course groups, then an edge is added to the learning course graph, and the weight of this edge is increased using an association weight coefficient. If some courses do not have associations in the learning course graph, then associations are established for these courses to construct a new learning course graph, which is the optimized learning course graph. The learning course graph of this invention is as follows: Figure 4 As shown, 0.25 represents the association weight coefficient. Since course C contains four edges, the association weight coefficient is 1 / 4. If two more edges are added between course C and course B, the association weight coefficient becomes 1 / 6, and the association weight coefficient between course C and course B becomes 3 / 6. The association weight coefficient is related to the number of edges between different nodes.

[0111] S3. Identify multiple target learner course sets, where each target learner course set includes multiple target learner courses, and each target learner course set corresponds one-to-one with a learner.

[0112] It should be explained that target learner courses refer to courses that are relevant to a specific learner (medical graduate student) and meet their learning needs and goals. These courses are determined based on the learner's professional direction, research interests, and career plans.

[0113] S4. Extract target learner course sets sequentially from multiple target learner course sets, and develop personalized learning course sets based on the target learner course sets.

[0114] In detail, the development of personalized learning course sets based on the target learners' course sets includes:

[0115] Undergraduate majors are obtained from the target learner course set, and course grade groups are obtained from the target learner course set, wherein the course grade groups in the course grade group set correspond one-to-one with the learner courses in the target learner course set.

[0116] The course grade sets are classified into basic course grade sets, professional course grade sets, and practical course grade sets. The highest and lowest basic course grades are calculated based on the basic course grade sets.

[0117] The highest and lowest professional course grades are calculated based on the professional course grade set, and the highest and lowest practical course grades are calculated based on the practical course grade set.

[0118] The strengths and weaknesses of courses are determined based on the highest and lowest grades in basic courses, the highest and lowest grades in professional courses, the highest and lowest grades in practical courses, and the lowest grades in practical courses.

[0119] Professional data is obtained based on undergraduate majors and research directions. Cross-disciplinary judgment is performed on the professional data to obtain judgment data, which can be either cross-disciplinary or non-cross-disciplinary data.

[0120] If the data is determined to be cross-disciplinary data, then obtain the professional span value and determine whether the professional span value is greater than the preset professional span threshold. If the professional span value is greater than the preset professional span threshold, then develop a cross-disciplinary learning course set based on the advantageous and disadvantageous courses.

[0121] If the data is determined to be non-interdisciplinary, then a non-interdisciplinary learning course set is developed based on the strengths and weaknesses of the courses.

[0122] A set of interdisciplinary or non-interdisciplinary learning courses can be used as a personalized learning course set, which includes one or more personalized learning courses.

[0123] It should be explained that the target learner's course set refers to the collection of all courses taken by a learner. Undergraduate major refers to the learner's undergraduate major. Research direction refers to the learner's current or future research direction. Basic course grade set refers to the learner's grades in basic courses. Professional course grade set refers to the learner's grades in professional courses. Practical course grade set refers to the learner's grades in practical courses. Highest and lowest basic course grades refer to the highest and lowest grades in the basic course grade set, respectively. Highest and lowest professional course grades refer to the highest and lowest grades in the professional course grade set, respectively. Highest and lowest practical course grades refer to the highest and lowest grades in the practical course grade set. Advantageous courses are those where the learner's grades are higher than other courses in a certain category. Disadvantageous courses are those where the learner's grades are the lowest among all courses. Professional data refers to data related to the learner's undergraduate major and research direction. For example, learner A's professional data = {Undergraduate major: Medicine, Research direction: Cardiovascular disease research}. The acquisition of the major span value refers to using a quantitative method (such as major similarity calculation) to measure the difference between the undergraduate major and the research direction. The TF-IDF algorithm is used to extract the undergraduate major name and research major name from the undergraduate major and research direction respectively. A pre-trained Word2Vec model is used to convert the extracted undergraduate major name and research major name into numerical word vectors for undergraduate major name and research major name respectively. Cosine similarity is used to calculate the cosine similarity between the undergraduate major name and research direction word vectors. The cosine similarity represents the degree of similarity between the undergraduate major and research direction; the larger the cosine similarity, the more similar the undergraduate major and research direction are, and the smaller the major span. To ensure a positive correlation between the major span value and cosine similarity, the reciprocal of the cosine similarity value is taken as the major span value. The larger the reciprocal value, the smaller the cosine similarity, indicating a greater difference between the undergraduate major and research direction. The major span threshold is a pre-set value used to determine whether the major span is less than the major span threshold. The method of developing an interdisciplinary learning curriculum based on strengths and weaknesses refers to selecting courses related to weaknesses, combined with the learner's undergraduate major and research direction, to help them fill knowledge gaps. It also involves selecting advanced or interdisciplinary courses related to strengths to further enhance their knowledge in their areas of expertise, and choosing courses that facilitate a smooth transition from their undergraduate major to their graduate research direction. Therefore, it is necessary to supplement foundational knowledge, bridging courses, and advanced courses in the target field. The method for developing a non-interdisciplinary learning curriculum based on strengths and weaknesses is the same as that for developing an interdisciplinary curriculum based on strengths and weaknesses, and will not be repeated here. A non-interdisciplinary learning curriculum refers to a course plan designed for learners from non-interdisciplinary backgrounds.The interdisciplinary and non-interdisciplinary learning course sets described in this embodiment of the invention each include one or more learning courses. Personalized learning courses refer to individualized course plans developed based on the learner's strengths, weaknesses, and interdisciplinary / non-interdisciplinary learning strategies. For example, if a student performs well in basic medical courses but poorly in clinical practice courses, it indicates that the student may have a strong grasp of theoretical knowledge but needs to improve practical skills, requiring supplementary courses in clinical practice and practical operations. Course matching degree refers to calculating the degree of match between the learner and a particular course based on the learner's personalized learning needs. The steps for quantifying the difficulty value of personalized learning courses in this embodiment of the invention are as follows: Several students who have studied the same personalized learning course for the same duration are selected. These students are determined by the operator and must be representative. A personalized learning course exam is used to test the difficulty of the course for these students. The exam scores are sorted in descending order, and the top A% and bottom A% of exam scores are extracted. A% is a preset extraction percentage by the operator. 100 is used as the maximum score. The scores obtained after elimination are extracted, and 100 is taken as the maximum difficulty value of the personalized learning course. The difficulty value of each exam score is obtained by subtracting the scores obtained after elimination from 100. All the obtained difficulty values ​​are summed to obtain a set of personalized learning course difficulty values. The set of personalized learning course difficulty values ​​is averaged to obtain the average value, which is used as the average personalized course difficulty value. The process of developing an interdisciplinary learning course set based on advantageous and disadvantageous courses involves acquiring all knowledge points from both advantageous and disadvantageous courses, identifying courses that contain 60% of the knowledge points from advantageous courses and 80% of the knowledge points from disadvantageous courses, and designating these courses as interdisciplinary learning courses. This process is then used to compile the interdisciplinary learning courses into a course set. S5: Receive medical course optimization instructions and, based on these instructions, the optimized learning course map, and the personalized learning course set, obtain the optimal recommended course sequence.

[0124] Specifically, the optimal recommended course sequence includes three optimal recommended courses.

[0125] In detail, obtaining the optimal recommended course sequence based on medical course optimization instructions, optimized learning course map, and personalized learning course set includes:

[0126] According to the medical course optimization instructions, courses to be recommended are extracted sequentially from the optimized learning course map. The course matching degree is calculated based on the personalized learning course set and the courses to be recommended. The course matching degree is then compared with the preset course matching degree threshold.

[0127] If the course matching degree is greater than the preset course matching degree threshold, then the course to be recommended will be used as the recommended course.

[0128] If the course matching degree is less than or equal to the preset course matching degree threshold, then return to the step of sequentially extracting recommended courses from the optimized learning course graph until all recommended courses in the optimized learning course graph have been extracted.

[0129] Summarize the recommended courses to obtain recommended course groups, and obtain the optimal recommended course sequence based on the recommended course groups.

[0130] It should be explained that the medical course optimization command is an instruction used to trigger the course recommendation process. Courses to be recommended refer to the learning courses extracted sequentially from the optimized learning course graph. The course matching threshold is a pre-set value used to determine whether the courses to be recommended match the learner's personalized learning course set. Recommended courses are those with a matching degree higher than the course matching threshold. The recommended course group is the set of all recommended courses. Obtaining the optimal recommended course sequence based on the recommended course group means extracting the three recommended courses with the highest matching degree from the recommended course group and sorting them to obtain the optimal recommended course sequence. Extracting courses to be recommended sequentially from the optimized learning course graph according to the medical course optimization command means that when the system receives the optimization command, it initiates an automated process to extract courses from the optimized learning course graph and filter and recommend them based on the learner's personalized needs.

[0131] In detail, the calculation of course matching degree based on the personalized learning course set and the courses to be recommended includes:

[0132] Extract personalized learning courses from the personalized learning course collection, obtain personalized course knowledge point groups from the personalized learning courses, summarize the personalized course knowledge point groups, and obtain a personalized course knowledge point group set.

[0133] Obtain the knowledge point set of the courses to be recommended, calculate the overlap ratio of knowledge points based on the knowledge point set and the knowledge point group set of personalized courses, analyze the difficulty of the courses to be recommended, and obtain the difficulty value of the courses to be recommended.

[0134] The difficulty of each personalized learning course in the personalized learning course set is analyzed to obtain a set of personalized learning course difficulty values. The average personalized course difficulty value is then calculated based on the set of personalized learning course difficulty values.

[0135] The course difficulty difference is calculated based on the difficulty value of the course to be recommended and the average personalized course difficulty value. The course difficulty difference is the absolute difference between the difficulty value of the course to be recommended and the average personalized course difficulty value.

[0136] Using a pre-built NLP model, vector transformation is performed on both the recommended courses and the personalized learning course set to obtain the embedded vectors of the recommended courses and the set of embedded vectors of the personalized courses. In the set of embedded vectors of the personalized courses, there is a one-to-one correspondence between the embedded vectors of the personalized courses and the personalized learning courses.

[0137] The average semantic distance is calculated based on the embedding vectors of the courses to be recommended and the set of embedding vectors of personalized courses. The course synergy effect value between the courses to be recommended and the set of personalized learning courses is calculated based on the average semantic distance, the courses to be recommended and the set of personalized learning courses, and the course moderating factor is obtained.

[0138] The course matching degree is calculated based on the overlap ratio of knowledge points, the difference in course difficulty, the course synergy effect value, and the course adjustment factor.

[0139] It should be explained that the "personalized course knowledge point group" for obtaining personalized learning courses refers to extracting knowledge points from course content (such as course syllabus, introduction, and teaching objectives). In this embodiment of the invention, the knowledge points refer to core concepts, theories, formulas, theorems, and professional terms in the course content, extracted from the personalized learning course or the course to be recommended using natural language processing technology. Knowledge point overlap means that shared knowledge points in two courses represent the same meaning. For example, if the knowledge points in the personalized learning course include: {cardiovascular system, pathology, cell division}, and the knowledge points in the course to be recommended include: {cardiovascular system, pathology, genetic variation}, then the knowledge points overlap between the personalized learning course and the course to be recommended are {cardiovascular system, pathology}. The knowledge point set of the course to be recommended refers to the set of all knowledge points covered in the course to be recommended. The personalized course knowledge point group refers to the set of knowledge points extracted from the personalized learning course. The personalized course knowledge point group set refers to the set of all personalized course knowledge point groups. The calculation of the knowledge point overlap ratio based on the knowledge point set and the personalized course knowledge point group set refers to obtaining all personalized course knowledge points from the personalized course knowledge point group set, obtaining all knowledge points of the course to be recommended from the knowledge point set, obtaining an overlapping knowledge point set from all personalized course knowledge points and all knowledge points of the course to be recommended, and dividing the overlapping knowledge point set by all knowledge points of the course to be recommended to obtain the knowledge point overlap ratio. The overlapping knowledge point set refers to the intersection of all personalized course knowledge points and all knowledge points of the course to be recommended. The course difficulty analysis of the course to be recommended refers to using the Analytic Hierarchy Process (AHP) to analyze the course difficulty of the course to be recommended. The method for analyzing the course difficulty of each personalized learning course in the personalized learning course set is the same as the method for analyzing the course difficulty of the course to be recommended, and will not be repeated here. The difficulty value of the course to be recommended refers to the quantitative value of the course difficulty obtained after analysis. The personalized learning course difficulty value set refers to the set of difficulty values ​​of all personalized learning courses. The personalized learning course difficulty value refers to the quantitative value of the personalized learning course difficulty obtained after analysis.

[0140] Importantly, the average personalized course difficulty value refers to the average of the set of personalized learning course difficulty values. An NLP model is a pre-trained model used for natural language processing to convert text into embedding vectors, facilitating the calculation of semantic similarity. Examples of NLP models include BERT and Word2Vec. The embedding vector of the course to be recommended refers to the numerical vector obtained by converting the content of the course to be recommended using an NLP model. The set of personalized course embedding vectors refers to the set of numerical vectors obtained by converting the content of each personalized learning course in the personalized learning course set using an NLP model. The step of calculating the average semantic distance based on the embedding vectors of the courses to be recommended and the set of personalized course embedding vectors involves calculating the cosine similarity between the embedding vector of the course to be recommended and the embedding vector of each personalized course in the set of personalized course embedding vectors, obtaining a set of cosine similarities, calculating the average value of the cosine similarity set, and using the average value of the cosine similarity set as the average semantic distance. The cosine similarity set refers to the set composed of all cosine similarities. The formula for calculating the course synergy effect value between the courses to be recommended and the personalized learning courses in the step of calculating the course synergy effect value based on the average semantic distance, the courses to be recommended, and the set of personalized learning courses is as follows:

[0141]

[0142] Where T represents the course synergy effect value, exp(*) represents the exponential function, and H represents the number of personalized learning course difficulty levels. v represents the average semantic distance. c This represents the embedding vector of the course to be recommended. The Cth element in the set of embedded vectors representing personalized courses o A personalized course is embedded in a vector group, C o This represents the index of the personalized course embedding vector in the personalized course embedding vector group, where || ||2 represents the modulo length.

[0143] Specifically, obtaining the course adjustment factor includes:

[0144] Obtain the knowledge point sets of the advantageous courses and the disadvantageous courses. Calculate the similarity of the first course based on the knowledge point sets of the advantageous courses and the similarity of the second course based on the knowledge point sets of the disadvantageous courses.

[0145] The first course similarity is obtained by mapping the similarity using a pre-constructed logical function and a preset molecular adjustment factor;

[0146] The second course similarity is mapped using the aforementioned logical function to obtain the second mapping similarity. The course adjustment factor is calculated based on the first mapping similarity and the second mapping similarity, wherein the course adjustment factor is the product of the first mapping similarity and the second mapping similarity.

[0147] It should be explained that the method for obtaining the knowledge point sets of advantageous and disadvantageous courses is the same as the method for obtaining the knowledge point set of the course to be recommended and the personalized course knowledge point group of the personalized learning course, and will not be repeated here. The knowledge point set of advantageous courses refers to the set of knowledge points extracted from advantageous courses. The knowledge point set of disadvantageous courses refers to the set of knowledge points extracted from disadvantageous courses. Calculating the similarity of the first course based on the knowledge point set and the knowledge point set of advantageous courses means calculating the similarity of the first course using the Jaccard similarity calculation formula. Calculating the similarity of the second course based on the knowledge point set and the knowledge point set of disadvantageous courses means calculating the similarity of the second course using the Jaccard similarity calculation formula. The first course similarity refers to the similarity between the course to be recommended and the learner's advantageous courses. The second course similarity refers to the similarity between the course to be recommended and the learner's disadvantageous courses.

[0148] It should be explained that a logistic function is a function that maps input values ​​to a specific output range, used to adjust similarity values. Mapping similarity values ​​to a specific range (e.g., 0 to 1) better reflects their importance. The first mapped similarity refers to the similarity after mapping the first course similarity using the logistic function and a numerator adjustment factor. The second mapped similarity refers to the similarity after mapping the second course similarity using the logistic function. The numerator adjustment factor is a pre-set value. For example, the numerator adjustment factor is 2. The numerator adjustment factor is used to expand the adjustment range of advantageous courses from [0, 1] to [0, 2], thereby achieving a reasonable enhancement of matching at high similarity, rather than simply maintaining neutrality. Course adjustment factors are used to adjust course matching, ensuring that recommended courses not only match the learner's advantageous courses but also help improve the learning effect of disadvantaged courses.

[0149] In detail, the formula for calculating the course matching degree is as follows:

[0150]

[0151] Where M represents the course matching degree, The Cth value represents the set of difficulty values ​​for personalized learning courses. r Each personalized learning course difficulty value, H represents the number of personalized learning course difficulty values ​​in the set, and d c Z represents the difficulty level of the course to be recommended, T represents the overlap ratio of knowledge points, Q represents the synergistic effect of the course, and Q represents the course adjustment factor. C represents the difference in course difficulty. r This represents an index representing the difficulty values ​​of personalized learning courses.

[0152] It should be explained that the course synergy value refers to the synergistic effect between the recommended course and the personalized learning course, that is, their complementarity in the learning path. Course matching degree refers to the degree of match between the recommended course and the learner, calculated after comprehensively considering the overlap of knowledge points, the difference in course difficulty, the course synergy value, and the course adjustment factor. If the course matching degree is too low, it means that the course is not suitable for the learner, the gap is too large, and it is not suitable for the learner's learning. Therefore, a tiered learning course should be set up, where learners first learn simpler courses and then more difficult courses, making it easier later. The course difficulty value is calculated from the scores obtained by different learners in the course examination and is used to assess the difficulty level of different learners when learning the course, that is, it can characterize the universality of the course and the learner. Different learners have varying levels of understanding or perception of different courses. Therefore, when recommending courses to learners, it is necessary to consider not only the general difficulty of the course but also the suitability of the course to the learner. This includes factors such as the overlap of knowledge points, the difference in difficulty between courses, the synergistic effect of courses, and the course adjustment factor. By considering these factors, the matching degree between the course and different learners can be calculated, which reflects how well the course should suit the learner under objective conditions.

[0153] S6. Obtain the ideal learning courses for the target learners and determine whether the optimal recommended course sequence contains ideal learning courses.

[0154] It should be explained that the ideal learning course refers to the course that the target learner (medical graduate student) expects to learn.

[0155] S7. If the optimal recommended course sequence does not contain an ideal learning course, then import the ideal learning course into the optimized learning course graph to obtain an updated learning course graph. Use the updated learning course graph as the optimized learning course graph and return to the step of obtaining the optimal recommended course sequence based on the medical course optimization instruction, the optimized learning course graph, and the personalized learning course set.

[0156] It should be explained that updating the learning course graph refers to importing the ideal learning course into the optimized learning course graph when the optimal recommended course sequence does not contain one, resulting in a new graph. The updated learning course graph will replace the original optimized learning course graph and be used in the next round of course optimization to ensure that course recommendations are more closely aligned with the needs of the target learners.

[0157] S8. Obtain the execution count of the optimal recommended course sequence obtained based on the medical course optimization instruction, the optimized learning course map, and the personalized learning course set. Compare the execution count with a preset execution count threshold. If the execution count is greater than the preset execution count threshold, then any one of the optimal recommended courses in the optimal recommended course sequence is taken as the ideal learning course.

[0158] It should be explained that the execution count is used to determine whether the system needs to stop the optimization process to avoid infinite loops. The execution count threshold is a pre-set value used to limit the maximum number of times the system performs optimization operations.

[0159] S9. Summarize the ideal learning courses to obtain the ideal learning course set, and optimize the postgraduate medical courses based on semantic reasoning based on the ideal learning course set.

[0160] It should be explained that the ideal learning curriculum set refers to the collection of all the ideal learning courses for the target learner.

[0161] To address the problems described in the background section, this invention identifies a medical graduate student database and obtains a set of medical learning courses based on this database. This set of medical learning courses includes multiple medical courses. By identifying the medical graduate student database, this invention integrates scattered student learning information and course information into a unified data source. A learning course graph is constructed based on the medical learning course set and then optimized to obtain an optimized learning course graph. This invention's learning course graph graphically presents the semantic relationships between medical learning courses, making the complex relationships between courses intuitive and easy to understand. Through optimization of the learning course graph, potential semantic connections between courses can be uncovered. This invention discovers previously unnoticed course combinations or knowledge expansion paths, providing new insights for course optimization and adjustment. It identifies multiple target learner course sets, each containing multiple target learner courses, with a one-to-one correspondence between the target learner course set and the learner. This ensures that each target learner course set accurately reflects the courses each learner has already studied or plans to study, providing a foundation for developing personalized learning course sets for each learner. The invention sequentially extracts target learner course sets from multiple target learner course sets and develops personalized learning course sets based on these sets. This invention develops personalized learning course sets based on each learner's target learner course set, enabling... By fully considering learners' existing knowledge and learning goals, and tailoring suitable learning courses for them, this invention helps improve learners' learning interest and effectiveness, making the courses more aligned with their actual needs. Receiving medical course optimization instructions, the invention obtains an optimal recommended course sequence based on these instructions, an optimized learning course map, and a personalized learning course set. This optimal recommended course sequence includes three optimal recommended courses. By combining the optimized learning course map and the personalized learning course set, this invention obtains an optimal recommended course sequence, providing learners with course recommendations that best match their needs and the logic of the curriculum system. This improves the accuracy and effectiveness of course recommendations, identifies the ideal learning courses for the target learner, and determines whether an ideal optimal recommended course sequence exists. This invention, by determining whether an ideal learning course exists for the target learner within the optimal recommended course sequence, ensures that the recommended courses meet the learner's personalized expectations. If no ideal learning course exists within the optimal recommended course sequence, the ideal learning course is imported into an optimized learning course graph to obtain an updated learning course graph. This updated learning course graph is then used as the optimized learning course graph, returning to the step of obtaining the optimal recommended course sequence based on the medical course optimization instructions, the optimized learning course graph, and the personalized learning course set. This invention, through continuous updating and optimization of the course graph, improves the accuracy and adaptability of course recommendations, ensuring that the recommended courses better meet the learner's expectations, continuously optimizing the course recommendation results, and increasing learner satisfaction.The invention obtains the execution count of the optimal recommended course sequence obtained based on medical course optimization instructions, optimized learning course graph, and personalized learning course set. The execution count is compared with a preset execution count threshold. This threshold prevents the algorithm from entering an infinite loop when an ideal learning course cannot be found, improving efficiency and stability. If the execution count exceeds the preset threshold, any optimal recommended course in the optimal recommended course sequence is selected as the ideal learning course. These ideal learning courses are then aggregated to obtain an ideal learning course set. Based on this set, semantic reasoning-based optimization of postgraduate medical courses is performed. This invention improves the quality and applicability of postgraduate medical courses, making them more aligned with students' learning needs and professional development requirements, thus contributing to the cultivation of more outstanding medical professionals. Therefore, this invention can improve the personalization and accuracy of course recommendations, promoting the development of medical education and talent cultivation.

[0162] like Figure 2 The diagram shown is a functional block diagram of a postgraduate medical course optimization system based on semantic reasoning provided in an embodiment of the present invention.

[0163] The semantic reasoning-based postgraduate medical course optimization system 100 of this invention can be installed in an electronic device. Depending on the functions implemented, the semantic reasoning-based postgraduate medical course optimization system 100 may include a course graph construction module 101, a personalized learning course formulation module 102, an optimal recommended course acquisition module 103, and a course optimization completion module 104. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and which are stored in the memory of the electronic device.

[0164] The course graph construction module 101 is used to identify the medical postgraduate database, obtain a medical learning course set based on the medical postgraduate database, wherein the medical learning course set includes multiple medical learning courses, construct a learning course graph based on the medical learning course set, and optimize the learning course graph to obtain an optimized learning course graph.

[0165] The personalized learning course development module 102 is used to identify multiple target learner course sets, wherein each target learner course set includes multiple target learner courses, and each target learner course set corresponds one-to-one with a learner. The target learner course sets are extracted sequentially from the multiple target learner course sets, and personalized learning course sets are developed based on the target learner course sets.

[0166] The optimal recommended course acquisition module 103 is used to receive medical course optimization instructions, acquire the optimal recommended course sequence based on the medical course optimization instructions, the optimized learning course map and the personalized learning course set, wherein the optimal recommended course sequence includes three optimal recommended courses, acquire the ideal learning course for the target learner, and determine whether there is an ideal learning course in the optimal recommended course sequence.

[0167] The course optimization completion module 104 is used to, if the optimal recommended course sequence does not have an ideal learning course, import the ideal learning course into the optimized learning course graph to obtain an updated learning course graph, use the updated learning course graph as the optimized learning course graph, return to the step of obtaining the optimal recommended course sequence based on the medical course optimization instruction, the optimized learning course graph, and the personalized learning course set, obtain the execution number of obtaining the optimal recommended course sequence based on the medical course optimization instruction, the optimized learning course graph, and the personalized learning course set, compare the execution number with a preset execution number threshold, if the execution number is greater than the preset execution number threshold, then any one of the optimal recommended courses in the optimal recommended course sequence is taken as the ideal learning course, the ideal learning courses are summarized to obtain the ideal learning course set, and the semantic reasoning-based optimization of postgraduate medical courses is completed based on the ideal learning course set.

[0168] In detail, the modules in the semantic reasoning-based postgraduate medical course optimization system 100 described in this embodiment of the invention employ the same methods as described above. Figure 1 The method used here is the same as the semantic reasoning-based optimization method for postgraduate medical courses described above, and it can produce the same technical effect, so it will not be repeated here.

[0169] like Figure 3 The diagram shown is a schematic representation of an electronic device for implementing a semantic reasoning-based method for optimizing postgraduate medical courses, according to an embodiment of the present invention.

[0170] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a semantic reasoning-based method for optimizing postgraduate medical courses.

[0171] The memory 11 includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a portable hard drive. In other embodiments, the memory 11 can be an external storage device of the electronic device 1, such as a plug-in portable hard drive, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 1. Furthermore, the memory 11 includes both internal storage units and external storage devices of the electronic device 1. The memory 11 can be used not only to store application software and various types of data installed on the electronic device 1, such as the code of a semantic reasoning-based postgraduate medical course optimization method program, but also to temporarily store data that has been output or will be output.

[0172] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., a semantic reasoning-based method for optimizing postgraduate medical courses) and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0173] The bus 12 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize the connection and communication between the memory 11 and at least one processor 10, etc.

[0174] Figure 3 Only electronic devices with components are shown; those skilled in the art will understand that... Figure 3The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0175] For example, although not shown, the electronic device 1 may also include a power supply (such as a battery) to power various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0176] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), which is typically used to establish communication connections between the electronic device 1 and other electronic devices.

[0177] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), or a standard wired or wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device 1 and to display a visual user interface.

[0178] The semantic reasoning-based postgraduate medical course optimization method program stored in the memory 11 of the electronic device 1 is a combination of multiple instructions, which, when run in the processor 10, can achieve the following:

[0179] The medical graduate student database was identified, and a medical learning course set was obtained based on the medical graduate student database. The medical learning course set includes multiple medical learning courses.

[0180] A learning course graph is constructed based on the medical learning course set, and then optimized to obtain an optimized learning course graph.

[0181] Multiple target learner course sets were identified. Each target learner course set includes multiple target learner courses, and each target learner course set corresponds one-to-one with a learner.

[0182] Extract target learner course sets sequentially from multiple target learner course sets, and develop personalized learning course sets based on the target learner course sets;

[0183] Receive medical course optimization instructions, and obtain the optimal recommended course sequence based on the medical course optimization instructions, the optimized learning course map, and the personalized learning course set. The optimal recommended course sequence includes three optimal recommended courses.

[0184] Obtain the ideal learning courses for the target learners and determine whether the optimal recommended course sequence contains ideal learning courses.

[0185] If the optimal recommended course sequence does not contain an ideal learning course, then the ideal learning course is imported into the optimized learning course graph to obtain an updated learning course graph. The updated learning course graph is then used as the optimized learning course graph, and the process of obtaining the optimal recommended course sequence based on the medical course optimization instruction, the optimized learning course graph, and the personalized learning course set is returned.

[0186] The number of times the optimal recommended course sequence is obtained based on the medical course optimization instructions, the optimized learning course map, and the personalized learning course set is obtained, and the number of executions is compared with the preset execution number threshold.

[0187] If the number of executions exceeds the preset execution threshold, then any one of the best recommended courses in the optimal recommended course sequence will be taken as the ideal learning course.

[0188] The ideal learning courses are compiled to obtain the ideal learning course set, and the postgraduate medical courses are optimized based on semantic reasoning based on the ideal learning course set.

[0189] Specifically, the processor 10's implementation method for the above instructions can be found in [reference needed]. Figures 1 to 4 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0190] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0191] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following:

[0192] The medical graduate student database was identified, and a medical learning course set was obtained based on the medical graduate student database. The medical learning course set includes multiple medical learning courses.

[0193] A learning course graph is constructed based on the medical learning course set, and then optimized to obtain an optimized learning course graph.

[0194] Multiple target learner course sets were identified. Each target learner course set includes multiple target learner courses, and each target learner course set corresponds one-to-one with a learner.

[0195] Extract target learner course sets sequentially from multiple target learner course sets, and develop personalized learning course sets based on the target learner course sets;

[0196] Receive medical course optimization instructions, and obtain the optimal recommended course sequence based on the medical course optimization instructions, the optimized learning course map, and the personalized learning course set. The optimal recommended course sequence includes three optimal recommended courses.

[0197] Obtain the ideal learning courses for the target learners and determine whether the optimal recommended course sequence contains ideal learning courses.

[0198] If the optimal recommended course sequence does not contain an ideal learning course, then the ideal learning course is imported into the optimized learning course graph to obtain an updated learning course graph. The updated learning course graph is then used as the optimized learning course graph, and the process of obtaining the optimal recommended course sequence based on the medical course optimization instruction, the optimized learning course graph, and the personalized learning course set is returned.

[0199] The number of times the optimal recommended course sequence is obtained based on the medical course optimization instructions, the optimized learning course map, and the personalized learning course set is obtained, and the number of executions is compared with the preset execution number threshold.

[0200] If the number of executions exceeds the preset execution threshold, then any one of the best recommended courses in the optimal recommended course sequence will be taken as the ideal learning course.

[0201] The ideal learning courses are compiled to obtain the ideal learning course set, and the postgraduate medical courses are optimized based on semantic reasoning based on the ideal learning course set.

[0202] In the embodiments provided by this invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and actual implementations may have other classification methods.

[0203] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0204] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0205] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0206] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for optimizing postgraduate medical courses based on semantic reasoning, characterized in that, The method includes: The medical graduate student database was identified, and a medical learning course set was obtained based on the medical graduate student database. The medical learning course set includes multiple medical learning courses. A learning course graph is constructed based on the medical learning course set, and then optimized to obtain an optimized learning course graph. The optimization of the learning course graph to obtain an optimized learning course graph includes: Perform the following operations on each course in the learning course map: Retrieve course textbooks based on the course content, retrieve the research major name based on the course textbooks, retrieve the set of university names based on the research major name, and then retrieve a set of similar textbooks with the same research major name from the set of university names. If the set of similar learning textbooks is empty, then the research major name is expanded to obtain an updated major name, and the updated major name is used as the research major name. Then, the step of retrieving the set of school names based on the research major name is returned until the set of similar learning textbooks is not empty. If the set of similar learning textbooks is not empty, then a set of similar learning courses is obtained based on the set of similar learning textbooks, and each similar learning course in the set of similar learning courses is combined with the learning course to obtain an associated similar course group. Summarize and associate similar course groups to obtain a set of similar course groups, and construct and optimize the learning course graph based on the set of similar course groups and the learning course graph. Multiple target learner course sets were identified. Each target learner course set includes multiple target learner courses, and each target learner course set corresponds one-to-one with a learner. Extract target learner course sets sequentially from multiple target learner course sets, and develop personalized learning course sets based on the target learner course sets; Receive medical course optimization instructions, and obtain the optimal recommended course sequence based on the medical course optimization instructions, the optimized learning course map, and the personalized learning course set. The optimal recommended course sequence includes three optimal recommended courses. The step of obtaining the optimal recommended course sequence based on medical course optimization instructions, optimized learning course map, and personalized learning course set includes: According to the medical course optimization instructions, courses to be recommended are extracted sequentially from the optimized learning course map. The course matching degree is calculated based on the personalized learning course set and the courses to be recommended. The course matching degree is then compared with the preset course matching degree threshold. If the course matching degree is greater than the preset course matching degree threshold, then the course to be recommended will be used as the recommended course. If the course matching degree is less than or equal to the preset course matching degree threshold, then return to the step of sequentially extracting recommended courses from the optimized learning course graph until all recommended courses in the optimized learning course graph have been extracted. Summarize recommended courses to obtain recommended course groups, and obtain the optimal recommended course sequence based on the recommended course groups; The calculation of course matching degree based on the personalized learning course set and the courses to be recommended includes: Extract personalized learning courses from the personalized learning course collection, obtain personalized course knowledge point groups from the personalized learning courses, summarize the personalized course knowledge point groups, and obtain a personalized course knowledge point group set. Obtain the knowledge point set of the courses to be recommended, calculate the overlap ratio of knowledge points based on the knowledge point set and the knowledge point group set of personalized courses, analyze the difficulty of the courses to be recommended, and obtain the difficulty value of the courses to be recommended. The difficulty of each personalized learning course in the personalized learning course set is analyzed to obtain a set of personalized learning course difficulty values. The average personalized course difficulty value is then calculated based on the set of personalized learning course difficulty values. The course difficulty difference is calculated based on the difficulty value of the course to be recommended and the average personalized course difficulty value. The course difficulty difference is the absolute difference between the difficulty value of the course to be recommended and the average personalized course difficulty value. Using a pre-built NLP model, vector transformation is performed on both the recommended courses and the personalized learning course set to obtain the embedded vectors of the recommended courses and the set of embedded vectors of the personalized courses. In the set of embedded vectors of the personalized courses, there is a one-to-one correspondence between the embedded vectors of the personalized courses and the personalized learning courses. The average semantic distance is calculated based on the embedding vectors of the courses to be recommended and the set of embedding vectors of personalized courses. The course synergy effect value between the courses to be recommended and the set of personalized learning courses is calculated based on the average semantic distance, the courses to be recommended and the set of personalized learning courses, and the course moderating factor is obtained. The course matching degree is calculated based on the overlap ratio of knowledge points, the difference in course difficulty, the course synergy effect value, and the course adjustment factor. The acquisition of course adjustment factors includes: Obtain the knowledge point sets of the advantageous courses and the disadvantageous courses. Calculate the similarity of the first course based on the knowledge point sets of the advantageous courses and the similarity of the second course based on the knowledge point sets of the disadvantageous courses. The first course similarity is obtained by mapping the similarity using a pre-constructed logical function and a preset molecular adjustment factor; The second course similarity is mapped using the aforementioned logical function to obtain the second mapping similarity. The course adjustment factor is calculated based on the first mapping similarity and the second mapping similarity, wherein the course adjustment factor is the product of the first mapping similarity and the second mapping similarity. Obtain the ideal learning courses for the target learners and determine whether the optimal recommended course sequence contains ideal learning courses. If the optimal recommended course sequence does not contain an ideal learning course, then the ideal learning course is imported into the optimized learning course graph to obtain an updated learning course graph. The updated learning course graph is then used as the optimized learning course graph, and the process of obtaining the optimal recommended course sequence based on the medical course optimization instruction, the optimized learning course graph, and the personalized learning course set is returned. The number of times the optimal recommended course sequence is obtained based on the medical course optimization instructions, the optimized learning course map, and the personalized learning course set is obtained, and the number of executions is compared with the preset execution number threshold. If the number of executions exceeds the preset execution threshold, then any one of the best recommended courses in the optimal recommended course sequence will be taken as the ideal learning course. The ideal learning courses are compiled to obtain the ideal learning course set, and the postgraduate medical courses are optimized based on semantic reasoning based on the ideal learning course set.

2. The method for optimizing postgraduate medical courses based on semantic reasoning as described in claim 1, characterized in that, The construction of the learning course map based on the medical learning course set includes: Extract one medical learning course from the medical learning course set one by one, and perform the following operations on each extracted medical learning course: Remove one of the extracted medical learning courses from the medical learning course set to obtain a simplified learning course set; Extract one simplified learning course from the simplified learning course set in sequence. Based on the extracted medical learning courses and the extracted simplified learning courses, obtain the list of medical course learners and the list of simplified course learners. Based on the list of medical course learners and the list of simplified course learners, confirm the list of common learners and the list of all learners. Course similarity is calculated based on the list of common learners and the list of all learners, where course similarity is the ratio of the number of common learners in the list of common learners to the number of all learners in the list of all learners. The course rating sets and simplified course rating sets corresponding to medical learning courses are collected respectively. The average course rating and average simplified course rating are calculated based on the course rating sets and simplified course rating sets. The modified cosine similarity is calculated based on the average course rating, the average simplified course rating, the course rating set, and the simplified course rating set. The comprehensive similarity is obtained based on course similarity and modified cosine similarity. The comprehensive similarity is then summarized to obtain a comprehensive similarity set. It is then determined whether there is a comprehensive similarity in the comprehensive similarity set that is greater than the preset comprehensive similarity threshold. If there is no comprehensive similarity greater than the preset comprehensive similarity threshold in the comprehensive similarity set, then return to the step of sequentially extracting medical learning courses from the medical learning course set until the medical learning course set is empty. If there is a comprehensive similarity greater than the preset comprehensive similarity threshold in the comprehensive similarity set, then the medical learning courses and simplified learning courses corresponding to the comprehensive similarity will be integrated to obtain an associated learning course group. By aggregating related learning course groups, a set of related learning course groups is obtained. A learning course graph is constructed based on the set of related learning course groups, which includes multiple learning courses.

3. The method for optimizing postgraduate medical courses based on semantic reasoning as described in claim 2, characterized in that, The formula for calculating the modified cosine similarity is as follows: in, Indicates the modified cosine similarity. This indicates the first course in the learning course assessment set. Each learning course rating, This indicates the first item in the simplified learning course rating set. A concise learning course rating. This represents the average course rating. This indicates the average rating of the simplified learning courses. This indicates the number of course ratings in the learning course rating set. This indicates the number of simplified learning course ratings in the simplified learning course rating set. An index representing the rating of the learning courses. An index representing the ratings for simplified learning courses.

4. The method for optimizing postgraduate medical courses based on semantic reasoning as described in claim 3, characterized in that, The process of developing personalized learning course sets based on the target learner's course set includes: Undergraduate majors are obtained from the target learner course set, and course grade groups are obtained from the target learner course set, wherein the course grade groups in the course grade group set correspond one-to-one with the learner courses in the target learner course set. The course grade sets are classified into basic course grade sets, professional course grade sets, and practical course grade sets. The highest and lowest basic course grades are calculated based on the basic course grade sets. The highest and lowest professional course grades are calculated based on the professional course grade set, and the highest and lowest practical course grades are calculated based on the practical course grade set. The strengths and weaknesses of courses are determined based on the highest and lowest grades in basic courses, the highest and lowest grades in professional courses, the highest and lowest grades in practical courses, and the lowest grades in practical courses. Professional data is obtained based on undergraduate majors and research directions. Cross-disciplinary judgment is performed on the professional data to obtain judgment data, which can be either cross-disciplinary or non-cross-disciplinary data. If the data is determined to be cross-disciplinary data, then obtain the professional span value and determine whether the professional span value is greater than the preset professional span threshold. If the professional span value is greater than the preset professional span threshold, then develop a cross-disciplinary learning course set based on the advantageous and disadvantageous courses. If the data is determined to be non-interdisciplinary, then a non-interdisciplinary learning course set is developed based on the strengths and weaknesses of the courses. A set of interdisciplinary or non-interdisciplinary learning courses can be used as a personalized learning course set, which includes one or more personalized learning courses.

5. The method for optimizing postgraduate medical courses based on semantic reasoning as described in claim 4, characterized in that, The formula for calculating the course matching degree is as follows: in, Indicates the course matching degree. The first in the set of difficulty values ​​for personalized learning courses Individualized learning course difficulty levels This indicates the number of personalized learning course difficulty values ​​within a set. This indicates the difficulty level of the course to be recommended. Indicates the percentage overlap of knowledge points. This represents the synergistic effect value of the courses. Indicates the course adjustment factor. This indicates the difference in course difficulty. This represents an index representing the difficulty values ​​of personalized learning courses.

6. A system applied to the semantic reasoning-based postgraduate medical course optimization method as described in claim 1, characterized in that, The system includes: The course graph construction module is used to identify the medical postgraduate database, obtain the medical learning course set based on the medical postgraduate database, where the medical learning course set includes multiple medical learning courses, construct a learning course graph based on the medical learning course set, and optimize the learning course graph to obtain an optimized learning course graph. The personalized learning course development module is used to identify multiple target learner course sets. Each target learner course set includes multiple target learner courses, and each target learner course set corresponds one-to-one with a learner. The target learner course sets are extracted sequentially from the multiple target learner course sets, and personalized learning course sets are developed based on the target learner course sets. The optimal recommended course acquisition module is used to receive medical course optimization instructions, obtain the optimal recommended course sequence based on the medical course optimization instructions, the optimized learning course map, and the personalized learning course set. The optimal recommended course sequence includes three optimal recommended courses. The module also obtains the ideal learning courses for the target learner and determines whether there are ideal learning courses in the optimal recommended course sequence. The course optimization completion module is used to import ideal learning courses into the optimized learning course graph if the optimal recommended course sequence does not contain ideal learning courses, thereby obtaining an updated learning course graph. This updated learning course graph is then used as the optimized learning course graph. The module returns to the step of obtaining the optimal recommended course sequence based on the medical course optimization instruction, the optimized learning course graph, and the personalized learning course set. It also retrieves the execution count of obtaining the optimal recommended course sequence based on the medical course optimization instruction, the optimized learning course graph, and the personalized learning course set, comparing the execution count with a preset execution count threshold. If the execution count is greater than the preset threshold, any one of the optimal recommended courses in the optimal recommended course sequence is taken as the ideal learning course. The ideal learning courses are then aggregated to obtain an ideal learning course set. Finally, the module completes the semantic reasoning-based optimization of postgraduate medical courses based on this ideal learning course set.

Images